Concept testing tells you whether people like an idea. Expert calls tell you whether the market will let it reach them — whether the channel will stock it, the regulator will permit it, and the incumbent will respond. Most failed launches fail on the second question having passed the first.
What does concept testing measure?
Appeal and comprehension among a target audience: whether people understand the idea and prefer it to alternatives presented alongside it. It does that well and it is a mature technique.
What it cannot measure is anything outside the respondent's head, which includes most of what determines whether a launch works.
It also measures preference among the alternatives you chose to show. A concept that tests well against two weak comparators has demonstrated something about the comparison rather than about the concept, and the selection of what to test against is made by the same team that wrote the concept. That is a known limitation of the method and it is rarely stated in the output.
| Concept testing | Expert calls | |
|---|---|---|
| Answers | Whether the target audience finds the idea appealing | Whether the market will let it reach them |
| Blind to | Channel, regulation, competitor response, procurement | Aggregate demand and preference at scale |
| Best used | After the market path is understood | Before committing to a concept |
What do expert calls add?
The market's constraints. Whether the channel will carry it, what shelf or slot it displaces, how long approval takes, and what the incumbent did last time someone attempted this — all answerable on a short run of expert calls.
None of that is knowable from a concept test at any sample size, because the respondents are not the ones who control it.
Expert conversations also establish who actually controls the decision, which a concept test assumes rather than discovers. In most B2B categories the person who would use the product and the person who would authorize buying it are different, and a test fielded to the first produces a confident reading of an opinion that does not decide anything.
Which question is the launch risk?
Usually the second. A concept that tests well and cannot reach the buyer fails; a concept that tests moderately and reaches the buyer easily frequently succeeds — the same reason willingness-to-pay research misses when it reaches the wrong person.
That asymmetry is why testing appeal first and feasibility later gets the sequence backward.
The asymmetry has a cost implication too. Feasibility work is usually cheaper than concept testing, because it needs a handful of informed conversations rather than a sample large enough to compare options. Running the cheaper thing second, after the more expensive thing has already been paid for, is a sequencing error in economic terms as well as methodological ones.
Does concept testing work in B2B?
With care. Sample sizes are small, the respondent is often not the budget holder, and the concept usually cannot be shown widely without disclosure risk, which limits what a survey can carry.
It is a consumer technique applied to a different buying structure, and the adaptation is rarely made explicit in the output.
Disclosure risk is the constraint teams most often underestimate. Showing a concept widely in a small market means showing it to people who talk to each other and, frequently, to competitors. That argues for testing with fewer, better-chosen respondents under a clearer agreement rather than for buying a larger sample and hoping the market is not that small.
How should you sequence them?
Establish the market path first, then test the concept within it. That way the concept being optimized is one that can actually be delivered — and bought, rather than piloted and never renewed.
Reversing the order produces a well-optimized proposition against constraints nobody had discovered yet.
The sequence also produces a better concept rather than only a better-informed decision. Knowing that the channel will carry a certain format, or that approval takes two quarters, changes what the proposition should be before anyone tests it — and a concept shaped by those constraints tends to test better than one that has to be adjusted afterward.